What is the relationship between a programming language and machine code?

Answer 1

About the relationship between engineering / architecture and Plumbing in building houses.

Computer Science looks at how one figures out how to develop algorithms and methods of solving a problem. Computer Programming is the implementation in software code of the algorithm.

Answer 2

Machine code is the only language native to the machine. But writing machine code programs is laboriously time-consuming and highly prone to error. To make it easier to produce machine code programs, we use programming languages. The machines won't understand these languages, of course, but they can be programmed to translate them into machine code. These translators are called assemblers, compilers and interpreters.

An assembler is the simplest translator. Assembly language is a second generation language (machine code is first generation) and is really nothing more than a symbolic version of machine code, where every operation code (opcode) maps to an easy-to-remember mnemonic such as MOV (move) or an acronym such JNZ (jump if not zero). Numeric operands can be entered using decimal notation as well as binary, but we can also use the more convenient hexadecimal and octal notations for binary data. We can also create user-defined names (variables) to refer to a memory address or to an array of addresses. While this makes it much easier to write and maintain the machine code, it's extremely low-level. Everything has to be written in minute detail in terms of the machine, is highly repetitive and must be heavily commented to keep track of what each section of code actually does.

Compilers and interpreters are complex translators. Both allow code to be written with a high-level of abstraction, hence compiled and interpreted languages are known as high-level languages. They are also known as third-generation languages because they can be used to create assembly language as well as machine code. The amount of abstraction varies from one language to the next, but they all make it much easier to encode ideas and concepts using high-level expressions that are much easier for humans to understand.

Compilers translate the entire source code into machine code while interpreters translate one statement at a time, execute it, then move onto the next. Interpretation results in much slower execution compared to a compiled machine code program, however a machine code program can only execute upon the machine type for which it was intended, which means the code must be recompiled separately for each supported platform. Interpreted programs, on the other hand, can be executed upon any machine with a suitable interpreter.

Some languages, such as Java, are both compiled and interpreted. That is, the Java compiler does not produce machine code it produces an intermediate code known as byte code which must then be interpreted by the Java virtual machine. Java programs execute much faster than a purely interpreted language, because the high-level statements are already translated into a low-level language that is much easier (and quicker) to interpret. Moreover, the same byte code can be executed upon any platform with a Java virtual machine implementation, which is pretty much everything these days.

While Java is the most popular language in use today, it is only suitable for writing applications software, it cannot be used to write low-level software such as drivers, operating system kernels or embedded systems.

C++ is the most widely-used general purpose compiled language, capable of writing code at both low and high level (hence it is often called an intermediate level language), but it can also interoperate with other high-level languages including C, Pascal and Fortran.